Taller 8

Stress-regulated grape CBFs H. Xiao et al.

Plant, Cell and Environment (2006) 29, 1410–1421

doi: 10.1111/j.1365-3040.2006.01524.x

Three grape CBF/DREB1 genes respond to low temperature, drought and abscisic acid HUOGEN XIAO*, MAHBUBA SIDDIQUA, SIOBHAN BRAYBROOK† & ANNETTE NASSUTH

Department of Molecular and Cellular Biology, University of Guelph, Guelph, Ontario, N1G 2W1, Canada

ABSTRACT The C-repeat (CRT)-binding factor/dehydrationresponsive element (DRE) binding protein 1 (CBF/ DREB1) transcription factors control an important pathway for increased freezing and drought tolerance in plants. Three CBF/DREB1-like genes, CBF 1–3, were isolated from both freezing-tolerant wild grape (Vitis riparia) and freezing-sensitive cultivated grape (Vitis vinifera). The deduced proteins in V. riparia are 63–70% identical to each other and 96–98% identical to the corresponding proteins in V. vinifera. All Vitis CBF proteins are 42–51% identical to AtCBF1 and contain CBF-specific amino acid motifs, supporting their identification as CBF proteins. Grape CBF sequences are unique in that they contain 20–29 additional amino acids and three serine stretches. Agro-infiltration experiments revealed that VrCBF1b localizes to the nucleus. VrCBF1a, VrCBF1b and VvCBF1 activated a green fluorescent protein (GFP) or glucuronidase (GUS) reporter gene behind CRT-containing promoters. Expression of the endogenous CBF genes was low at ambient temperature and enhanced upon low temperature (4 ∞C) treatment, first for CBF1, followed by CBF2, and about 2 d later by CBF3. No obvious significant difference was observed between V. riparia and V. vinifera genes. The expression levels of all three CBF genes were higher in young tissues than in older tissues. CBF1, 2 and 3 transcripts also accumulated in response to drought and exogenous abscisic acid (ABA) treatment, indicating that grape contains unique CBF genes. Key-words: abiotic stress; cold; cold acclimation; freezing and drought tolerance; Vitis.

INTRODUCTION Plants acclimate to environmental stresses, such as low temperature and drought, through biochemical and physiological processes that result from the induced expression or repression of a battery of target genes. One particular Correspondence: Annette Nassuth. Fax: +1 519 7671656; e-mail: [email protected] Present addresses: *Department of Food Science, University of Guelph, Guelph, Ontario, Canada; †Section of Plant Biology, University of California, Davis, CA, USA.

1410

regulatory sequence, C-repeat (CRT) or dehydrationresponsive element (DRE), was found in the promoter of target Arabidopsis thaliana genes whose expression contributes to both low temperature and drought tolerance (Stockinger, Gilmour & Thomashow 1997; Gilmour et al. 1998; Liu et al. 1998). This can be explained by the fact that tolerance to both stresses requires stability of cell components during dehydration. Two families of transcription factors binding to CRT/DRE have been identified in A. thaliana (Stockinger et al. 1997; Gilmour et al. 1998; Liu et al. 1998). The first, called CRT-binding factor/DRE binding protein 1 (CBF/DREB1), is encoded by genes whose expression is rapidly induced in response to low, non-freezing temperatures, but not by dehydration (Gilmour et al. 1998; Medina et al. 1999). The second family of transcription factors, called DRE binding protein 2 (DREB2), and one member of the first family, CBF4/DREB1D, are encoded by genes that are induced by drought, but not by low temperature stress (Liu et al. 1998; Haake et al. 2002). Although microarray and other analyses revealed that a number of regulatory pathways likely play a role in tolerance to low temperature and drought stress (Fowler & Thomashow 2002; Seki et al. 2002), the CBF/DREB1 pathway alone appears sufficient to increase stress tolerance. Constitutive expression of the CBF/DREB1 genes in transgenic A. thaliana plants induces expression from CRT/DREcontaining genes and results in an increase in freezing and drought tolerance without a prior stimulus (Jaglo-Ottosen et al. 1998; Liu et al. 1998; Kasuga et al. 1999; Gilmour et al. 2000; Haake et al. 2002). The potential utility of CBF/ DREB1 genes to improve stress tolerance prompted a search for similar genes in many other plants. CBF/ DREB1-encoding genes have now been identified in higher plants such as Brassica napus, tomato, wheat, rye, tobacco, cotton, capsicum, cherry, rice and barley (Jaglo et al. 2001; Choi, Rodriquez & Close 2002; Gao et al. 2002; Owens et al. 2002; Dubouzet et al. 2003; Xue 2003; Skinner et al. 2005). The encoded proteins were found to have several amino acid motifs in common. Putative functions have, so far, been proposed for only two: nuclear localization for the nuclear localization signal (NLS) motif and DNA binding for the AP2 domain. However, the motifs can be used to identify and classify members of this family of transcription factors (Jaglo et al. 2001; Owens et al. 2002; Al-daoud 2004; Skinner et al. 2005). For simplicity, we will refer to CBF/ DREB1 as CBF in the rest of this manuscript. © 2006 The Authors Journal compilation © 2006 Blackwell Publishing Ltd

Stress-regulated grape CBFs 1411 Vitis vinifera, the grape species that forms the basis of most major wine industries, is native to Asia Minor and is typically injured at temperatures below −20 °C (Winkler 1970). Freezing damage in winter and early spring results in severe losses in vineyards in the cool climate regions around the world. Different species of grape exhibit a broad range of cold hardiness, but little is known about freezing tolerance in Vitis at the molecular level. Here, we identify three CBF genes and their expression in grape in response to low temperature, drought and abscisic acid (ABA). Both V. vinifera and its hardy relative, Vitis riparia, were included to identify possible differences, which might suggest potential strategies to produce cultivated grapes with higher stress tolerance.

MATERIALS AND METHODS Plant materials and treatments Vitis riparia from Thunder Bay, Ontario, Canada and V. vinifera cv. Chardonnay were obtained from Chateau des Charmes Wineries in Niagara Peninsula, Ontario, Canada. Cuttings were rooted and maintained in the glasshouse or, where indicated, in controlled-environment growth chambers programmed for, starting at 0600 h, 16 h light at an intensity of 80–100 µm−2 s−1 and 22 °C followed by 8 h dark at 20 °C. Two-month-old plants cultured from the buds of V. riparia and V. vinifera were also used for some experiments. Tobacco (Nicotiana tabacum cv. Petite Havana or Nicotiana benthamiana) was grown in controlledenvironment growth chambers under the same conditions. Cold treatment was started by transferring the plants between 0900 and 1000 h to a growth chamber set at 4 °C and with constant light. Dehydration was induced by placing detached leaves on a dry filter paper. The final fresh weight (FW) of these leaves was 91.3% of the starting values after 15 min, 79.3% after 1 h and 22.9% after 24 h of desiccation. For ABA treatments, detached leaves were first soaked into 100 µM ABA in 0.02% v/v polyoxyethylene-sorbitan monolayrate (Tween-20) and then immediately placed on a filter paper soaked in ABA solution. As controls, detached leaves were placed on a filter paper soaked in water or Tween solution and kept humid. Treatments continued for various periods, as indicated in the text and in the figures. Collected samples, consisting of two to four young (first and second) or mature (fifth and sixth) leaves, young (green) or mature (brown) buds, apical tips and/or young stems (green stems starting just below the apical tip until the fourth leaf position), were immediately frozen in liquid N2 and stored at −80 °C until their use.

DNA extraction and digestion DNA from V. riparia Thunderbay, V. vinifera cv. Chardonnay and A. thaliana was extracted essentially according to the procedure described by Sambrook & Russell (2001). Aliquots of genomic DNA (10 µg) were digested overnight at 37 °C with the indicated restriction enzyme (Fermentas,

Burlington, ON, Canada), in the presence of 1% polyvinylpyrrolidone 40 (PVP40) to increase the DNA digestion (Nassuth, unpublished results).

Isolation of CBF-like genes from grape Partial putative CBF orthologs in V. riparia were amplified by PCR [using conditions described for reverse transcription (RT)–PCR, but without RT step and for 35 cycles]. Primers CBFd1-H (5′-TTYMRDGAGACDMGDC ACCC-3′) and CBFd4-C (5′-ARRAGMADNCCYTC NGCCAT-3′), designed based on the sequences encoding the conserved CBF-specific domains PKK/RPAGRxKFxETRHP and NMAEGMLLPP in A. thaliana CBF1, 2 and 3 (Gilmour et al. 1998; Medina et al. 1999) and Brassica CBF (Jaglo et al. 2001), were used. The resulting PCR products were cloned in pGEM-T easy vector (Promega, Madison, WI, USA) and sequenced using dye terminator cycle sequencing on an ABI PRISM model 377 (Guelph Molecular Supercentre, University of Guelph, Ontario, Canada). Flanking sequences were amplified from V. riparia genomic DNA by nested inverse PCR (iPCR) according to a protocol modified by Sambrook & Russell (2001) with two primer pairs, F1 and R1, and F2 and R2, specific for each obtained sequence (VrCBF1-F1: 5′-CCAGGATATG GCTAGGCACC-3′; VrCBF1-R1: 5′-CCGCACGCCTCT GTATATTG-3′; VrCBF1-F2: 5′-CTCAATTTCTCCGAC TCGGC-3′; VrCBF1-R2: 5′-GGTGTCGTGTCTCCCG GAA-3′; VrCBF2-F1: 5′-AACTTCTTCTGTCGTAG ACA-3′; VrCBF2-R1: 5′-TTATCCATTACGAAAGGT GT-3′; VrCBF2-F2: 5′-ATCGGTGGATAGTAAGAGC G-3′; VrCBF2-R2: 5′-AGATGACGATGATGGAGGTT3′; VrCBF3-F1: 5′-ATGCAGTGAAGACTCGCCTC-3′; VrCBF3-R1: 5′-GCTACAGGCAGTGACATGTG-3′; VrCBF3-F2: 5′-CTCTCCACATGGTTCGAGC-3′; VrCBF3R2: 5′-TAAGAAGAGGATGAAGACGG-3′). V. riparia genomic DNA was digested with either EcoRI, HindIII, XbaI and NcoI. The digested DNAs were purified using QIAEX II Kit (Qiagen, Valencia, CA, USA), and 300 ng was self-ligated overnight at 15 °C in a total volume of 100 µL with 0.04 U µL−1 T4 DNA ligase (Roche Diagnostics Canada, Laval, QC, Canada). Five microlitres of each ligation mixture was added directly to a 50 µL PCR reaction containing F1 and R1 primers. The PCR reaction conditions included a 3 min extension step. One microlitre of the firstround PCR mixture was added to a second-round PCR with the nested primers F2 and R2. Amplification was successful for HindIII- and XbaI-digested DNA with Vitis CBF1-specific primers (to give Vitis CBF1a and Vitis CBF1b sequence, respectively), for NcoI-digested DNA with Vitis CBF2-specific primers, and for XbaI-digested DNA with Vitis CBF3-specific primers. Amplified products were cloned into pGEM-T easy vector and sequenced. Primers designed against the obtained sequences were used to amplify the complete coding regions of the three V. riparia CBFs and their homologs in V. vinifera. These primers were VrCBF1-H-10 (5′-CTCTCTCTCCATGGAC TCGG-3′), VrCBF1-C930 (5′-TTAATTCTTCCTAATA

© 2006 The Authors Journal compilation © 2006 Blackwell Publishing Ltd, Plant, Cell and Environment, 29, 1410–1421

1412 H. Xiao et al. TAAGTATATATATTTATG-3′), VrCBF2-H-39 (5′-TCTC GTCTCCAACTCTTACT-3′), VrCBF2-C897 (5′-ACCT GAAGTCCATCCAAGTT-3′), VrCBF3-H-63 (5′-CTCT CAATCTCTTTCTACTTGC-3′) and VrCBF3-C802 (5′AATGTGAACACTAGGCAGTG-3′). At least four clones were sequenced for each Vitis CBF. The coding regions of the CBF grape genes described in this report have been deposited in the GenBank database under AY390370 (VrCBF1a), AY390371 (VrCBF1b), AY390373 (VrCBF2), AY390374 (VrCBF3), AY390372 (VvCBF1), AY390376 (VvCBF2) and AY390375 (VvCBF3).

Southern blot analysis Vitis riparia Thunderbay and V. vinifera cv. Chardonnay DNAs were digested with restriction enzymes whose recognition sites were absent in the determined Vitis CBF sequences. The digested DNA was separated by agarose gel electrophoresis, denatured and transferred to a positively charged nylon membrane (Roche) and bound using the Stratalinker 2400 UV cross linker (Stratagene, La Jolla, CA, USA). Digoxigenin-11-2′-deoxyuridine-5′-triphosphate (dUTP) (Roche)-labelled probes were prepared by PCR on the different grape CBF plasmid DNAs to produce fragments corresponding to either the complete CBF1 coding region (primers VrCBF1-H-10; VrCBF1-C753: 5′-TTAAT CATCATTCCACAAAGACAAGTCA-3′) or, for specific primers, the C-terminal region starting after AP2 of CBF1a, of CBF2 or of CBF3 (primers VrCBF1-H556: 5′-TG GATACTAAGAGGTCAGAG-3′ and VrCBF1-C930; VrCBF2-F1 and VrCBF2-C897; VrCBF3-F1 and VrCBF3C802). The procedures described in the digoxigenin (DIG) system user’s Guide for Filter Hybridization (Boehringer Mannhein/Roche) were followed for pre-hybridization, hybridization and development of the membrane. High sodium dodecyl sulphate (SDS) buffer [7% SDS, 50% deionized formamide, 5× standard saline citrate (SSC), 4% blocking reagent (Roche), 50 mM sodium phosphate (pH 7.0) and 0.1% N-laurolysarcosine] was used for both pre-hybridization for at least 1 h at 42 °C, and for hybridization overnight at 42 °C after the addition of 25 ng mL−1 heat-denatured DIG-labelled DNA probe.

Preparation of fusion protein construct for localization study Two fusion protein constructs were prepared for the nuclear localization experiment. For the Vitis CBF–green fluorescent protein (GFP) fusion, the coding sequence for V. riparia CBF1b was amplified by PCR from CBF1b clones using primers BamHI + VrCBF1-H1 (5′-GGCGGATC CAAGGAGATATAACAATGGACTCGGACCACGAA GAG-3′) and EcoRI + VrCBF1-C750 (5′-GGAATTCAT CATCATTCCACAAAGACAAGTCA-3′). Similarly, the coding sequence for modified green-fluorescent protein (mGFP5) (J. Haselhof, University of Cambridge, Cambridge, UK) was amplified without endoplasmic reticulum (ER)-targeting signal and ER- retention signal (KDEL), by

PCR from the small, cloning plasmid p35S–mgfp5 (Lee et al. 2001) using primers EcoRI + mgfp5-H3 (5′-GGAAT TCAGTAAAGGAGAAGAACTTTTC-3′) and SacI + mgfp5-C737 (5′-GGCGAGCTCTTATTTGTATAGTTCA TCCATGCC-3′). The BamHI–EcoRI-digested VrCBF1b and EcoRI–SacI-digested mgfp5 fragments were simultaneously ligated into BamHI–SacI-digested intermediate plasmid, and the resulting construct was sequenced to confirm that the construct encodes a 492 amino acid long CBF– GFP fusion protein. The VrCBF1b–mGFP5 fusion fragment was then subcloned into BamHI–SacI-digested pBI121 (Clontech, Mountain View, CA, USA) behind the cauliflower mosaic virus 35S (CaMV35S) promoter to produce pBIN35S::VrCBF1b–mGFP5. Similarly, part of the β-glucuronidase (GUS) coding sequence was amplified by PCR using primers BamH1 + gus-H1 (5′-GGCGG ATCCAAGGAGATATAACAATGTTACGTCCTGTAG AAAC-3′) and EcoRI + gus-C786 (5′-GGAATTCGACG CACAGTTCATAGAGAT-3′), digested and ligated with the mgfp5 fragment to produce a sequence encoding a GUS∆–GFP fusion protein similar in size (494 aa) to the VrCBF1b–mGFP5 protein. The resulting control plasmid was named pBIN35S::GUS∆–GFP. The vectors were introduced into Agrobacterium tumefaciens strain EHA105.

Preparation of transactivation constructs All constructs for transactivation experiments were first prepared in the smaller cloning plasmid p35S–mgfp5 or pGEM-T, to facilitate cloning and selection. Clones were confirmed by sequencing, and then introduced into pBI121 (Clontech). Both p35S–mgfp5 and pBI121 contain the same sequence from CaMV35S promoter to nopaline synthase (NOS) terminator (located on a HindIII–SacI fragment). Initial reporter plasmids contained mGFP5–ER, an ERtargeted, modified GFP. This reporter was chosen because ER localization apparently allows greater accumulation and better fluorescence detection of the GFP (Grebenok et al. 1997; Haseloff et al. 1997; Haseloff & Siemering 1998). mGFP5–ER, derived from p35S–mgfp5, was subcloned into pBI121 to produce pBIN35S::mGFP5–ER as a positive control. Three types of reporter plasmids were used: one with the A. thaliana RD29A promoter sequence, one with four CRT elements in front of the −46 minimal 35S promoter sequence and, as a negative control, one with the −46 minimal 35S promoter sequence only. The RD29A promoter sequence was amplified by PCR from A. thaliana genomic DNA with primers HindIII + RD29Apro-H-944 (5′-CCCAAGCTTGAGCCATAGATGCAATTC-3′) and BamHI + RD29Apro-H-22 (5′-CGGGATCCAATAGAA GTAATCAAACC-3′). The minimal 35S promoter sequence with or without 4CRT was amplified from vector DNA with primers HindIII + 4CRT 35S-H-46 (5′-CCGA AGCTTACCGACATTACCGACATTACCGACATTAC CGACATTACGCAAGACCCTTCCTCTA-3′) or HindIII + 35S-H-46 (5′-CCGAAGCTTCGCAAGACCCTTCCT CTA-3′) and BamHI/XbaI + 35S-C (5′-CGGGGATCCTC TAGAGTC-3′). The HindIII–BamHI fragments containing

© 2006 The Authors Journal compilation © 2006 Blackwell Publishing Ltd, Plant, Cell and Environment, 29, 1410–1421

Stress-regulated grape CBFs 1413 the promoters were used to replace the 35S promoter and produce, respectively, pBINRD29A::mGFP5–ER, pBIN4CRTmin35S::mGFP5-ER and pBINmin35S:: mGFP5-ER. For later experiments, where indicated, mGFP5-ER was replaced with GUSPlus (CAMBIA, Canberra, Australia) because GUSPlus contains an intron, thus allowing the detection of plant-specific glucuronidase expression only. The GUSPlus gene was amplified from pCAMBIA 1305.1 (GenBank AF364045; CAMBIA) with primers BamHI + GUSPlus-H1: 5′-GGCGGATCCAAG GAGATATAACAATGGTAGATCTGAGGGTAAATT TC-3′ and SacI + GUSPlus-C-2057: 5′-GGCGAGCTCA ATTCACACGTGATGGTGATG-3′, and subcloned to produce pBINmin35S::GUSPlus or pBIN4CRTmin35S:: GUSPlus. All constructs were introduced into A. tumefaciens strain EHA105. To prepare the effector plasmid, the coding sequence of A. thaliana CBF1 was amplified by PCR from A. thaliana genomic DNA using primers AtCBF1-H1 (5′-CGGGATCCTTATCCAGTTTCTTGAAA-3′) and AtCBF1–C766 (5′-CCATTCTAAAAAAGGAACTA-3′) and cloned into BamHI–SacI-digested plasmid to replace mGFP5–ER and produce pBIN35S::AtCBF1. Similarly, V. riparia CBF1a and CBF1b, and V. vinifera CBF1 were amplified from their plasmids with primers BamHI + VrCBF1-H1 and SacI + VrCBF1-C753 (5′-GGCGAGCT CTTAATCATCATTCCACAAAGACAAGTCA-3′) and subcloned to produce pBIN35S::VrCBF1a, pBIN35S:: VrCBF1b and pBIN35S::VvCBF1. All final constructs were confirmed by restriction digestion and introduced into A. tumefaciens strain EHA105 by the freeze–thaw method (Hofgen & Willmitzer 1988).

Agro-infiltration of tobacco leaves Agrobacterium containing individual constructs was cultured overnight at 28 °C and 300 r.p.m. in Luria Bertani medium containing 50 µg mL−1 of kanamycin and 10 µg mL−1 of rifampicin. One millilitre of the culture was transferred to 50 mL fresh medium containing 10 mM 2morpholinoethanesulfonic acid (MES) (pH 5.6) and 40 µM acetosyringone, and again cultured overnight until the OD600 was 1. The bacteria were pelleted at 2860 g in a Hermle Z320K centrifuge (Hermle, Gosheim, Germany) for 10 min and resuspended in 10 mM MgCl2 to give a final volume of 100 mL for nuclear localization experiments or 50 mL for transactivation experiments (i.e. equivalent to an OD of 0.5 or 1). Acetosyringone (150 µM) was added and the bacterial suspensions were kept at room temperature for at least 3 h without shaking. Expanded tobacco (N. tabacum cv. Petite Havana or N. benthamiana) leaves were infiltrated with a suspension of Agrobacterium using a 1 mL needleless syringe. For transactivation experiments, equal volumes of A. tumefaciens containing the reporter construct pRD29A::mGFP5– ER, and of A. tumefaciens containing an effector construct, such as 35S::AtCBF1, were mixed before infiltration (i.e. equivalent to a final OD of 0.5 for each A. tumefaciens strain). Controls with only one construct-

containing Agrobacterium were mixed with an equal volume of untransformed Agrobacterium. The infiltrated leaves were collected after 2–3 d for visualization of GFP fluorescence in epidermal peels (mounted in water) by epifluorescence microscopy (with a Carl Zeiss/Jenalumar SH250 UV illumination microscope, Jena, Germany) or confocal microscopy (with a Leica TCS SP2, Heidelberg, Germany), for GUS activity assay or, where indicated, for RNA isolation. The fluorometric analysis of GUS activity was carried out essentially as described by Jefferson 1987).

Semi-quantitative RT–PCR Total RNA was extracted from harvested leaves using the Plant RNeasy Mini Kit (Qiagen) according to Nassuth et al. (2000). Approximately 100 mg of plant tissue was ground in 2 mL of extraction lysis buffer, and RNA isolated from half of the extract was collected in 100 µL ribonuclease (RNase) free H2O. Total RNA (25 µL) was treated with 2 µL (4 U) of RNase-free deoxyribonuclease (DNase) (Ambion, Austin, TX, USA) in 1× DNase buffer [20 mM tris(hydroxymethyl)aminomethane (Tris)–HCl (pH 8.0) and 10 mM MgCl2] for approximately 1 h at 37 °C followed by incubation at room temperature for 2 min with 6 µL of DNase inactivation reagent (Ambion). The inactivation reagent was pelleted by centrifugation and aliquots of the RNA-containing supernatant were stored at −80 °C. One-tube RT–PCR (Nassuth et al. 2000) was used to semi-quantify specific mRNAs. Each RT–PCR mixture (final volume 25 µL) contained 10 mM Tris–HCl (pH 8.8), 100 mM KCl, 1.5 mM MgCl2, 200 µM deoxynucleotide triphosphates (dNTPs) (Mg balanced), 5 mM dithiothreitol (DTT), 2% (w/v) sucrose, 0.1 mM cresol red (Sigma, St. Louis, MO, USA), 0.2 U avian myeloblastosis virus reverse transcriptase (AMV RT) (Roche), 4 U Thermus aquaticus (Taq) polymerase (Fermentas), 2.5 µL of DNase-treated total RNA and 0.5 µM of each primer. The primers used were specific for Vitis CBF1 (VrCBF1-H464: 5′-AGTTG CAGACTCGAAGAAGG-3′ and VrCBF1-C778: 5′-AA TCTAAGCGCACCTATGTC-3′), Vitis CBF2 (VrCBF2-F1 and VrCBF2-C897), Vitis CBF3 (VrCBF3-F1 and VrCBF3C802), GFP (mGFP5–H11: 5′-GAGAAGAACTTTTCA CTGGA-3′ and mGFP5-C711: 5′-GTATAGTTCATCCA TGCCAT-3′ or mGFP5ER H34: 5′-TCACTTCTCCTAT CATTATCCTC-3′ and mGFP5ER C371: 5′-TCGTCCT TGAAGAAGATGGT-3′) or GUSPlus (GUSPlus-H-11 g: 5′-GGAGATATAACAATGGTAGATCTGAGG-3′ and GUSPlus-C599: 5′-GCGATTCATGCCATCACG-3′). The amplification was carried out in a thermocycler (PTC-100; MJ Research, Waltham, Germany) using one incubation at 42 °C for 45 min (RT) followed by 28–35 cycles at 94 °C for 30 s, 54 °C for 45 s and 72 °C for 1 min, with a final incubation at 72 °C for 5 min. The number of cycles (28–35) was chosen such that maximum amplification was not yet reached, as determined by gel electrophoresis. Identical reactions without RNA or reverse transcriptase were included for each set of RT–PCR as negative controls.

© 2006 The Authors Journal compilation © 2006 Blackwell Publishing Ltd, Plant, Cell and Environment, 29, 1410–1421

1414 H. Xiao et al.

Figure 1. Amino acid sequence alignment of CBF proteins from Vitis and Arabidopsis thaliana. The CBF sequences and their GenBank accession numbers are VrCBF1a (AY390370), VrCBF2 (AY390373) and VrCBF3 (AY390374) from Vitis riparia; VvCBF1 (AY390372), VvCBF2 (AY390376) and VvCBF3 (AY390375) from Vitis vinifera; and AtCBF1 (U77378) from A. thaliana. The dashes indicate gaps introduced for better alignment. Identical amino acids are shaded black, conserved substitutions are shaded grey. The total number of amino acids for each protein is indicated at the end of each sequence. The AP2/EREBP domain; putative nuclear localization signal (NLS); and A, P, DSAWRL and LWSY motifs are indicated by solid lines. Serine repeats are indicated by arrows.

Control reactions with primers specific for Vitis malate dehydrogenase (VvMDH-H968: 5′-GCATCTGTGGTT CTTGCAGG-3′ and VvMDH-C1163: 5′-CCCTTTGAG TCCACAAGCCAA-3′), tobacco malate dehydrogenase (NtMDH-H4: 5′-CCTGGTGTTGCCGCTGAT-3′ and NtMDH-C352: 5′-TTGCCCTAACAACATCAAGTGT-3′) or tobacco ribulose 1·5-bisphosphate carboxylase/oxygenase (Rubisco)-L (RbcL-H681: 5′-TGGACTTGATTTTAC CAAAGATGATG-3′ and RbcL-C1231: 5′-TGTCCTAAA GTTCCTCCACC-3′) were also included to show that equal amounts of RNA were used in each set of reactions.

RESULTS CBF genes are present in V. riparia and V. vinifera To identify and isolate CBF-like sequences from grape, we designed the degenerate primers CBFd1-H and CBFd4-C, based on conserved regions in A. thaliana and Brassica CBF proteins (Gilmour et al. 1998; Medina et al. 1999; Jaglo et al. 2001), and used these primers for PCR on V. riparia DNA. Two DNA fragments of about 500 and 400 bp were amplified, cloned and sequenced (data not shown). Three different sequences with high homology to A. thaliana CBF were identified in the 500 bp fragment-derived clones. Flanking sequences for each were amplified by nested iPCR reactions. Sequence type 1-specific primers amplified fragments of 2.7 and 2.2 kb on, respectively, HindIII- and XbaI- digested and self-ligated V. riparia genomic DNA. To confirm the obtained sequences, we designed primers outside the identified open reading frame and used those in PCR on undigested V. riparia DNA. Fourteen clones from four independent PCR reactions for the open reading frame were sequenced and it turned out that nine were identical to the HindIII fragment sequence, whereas the other five were identical to the XbaI fragment sequence. The two derived sequences were named, respectively, VrCBF1a and

VrCBF1b. Both contained an open reading frame for a protein of 251 amino acids with a predicted molecular mass of 27.6 kDa. VrCBF1a and VrCBF1b differ only at residues 20 and 108, with an asparagine and proline in VrCBF1a, and a serine and serine in VrCBF1b. One CBF1 sequence, VvCBF1, was also amplified from V. vinifera genomic DNA. The encoded protein was of similar size and 98% identical to VrCBF1a (Fig. 1). The grape CBF1 proteins are rich in serine (14–14.7%), alanine (9.2%), leucine (8.4%), arginine (6.8%) and aspartic acid (6.8%). The isoelectric point (pI) of VrCBF1a/1b and VvCBF1 are 8.01 and 8.9, respectively. Sequence type 2-specific primers amplified a 1.1 kb fragment on NcoI-digested and self-ligated genomic DNA. The subsequently amplified coding sequences of CBF2 from V. riparia and V. vinifera (Fig. 1) showed that VrCBF2 and VvCBF2 are 250 and 253 amino acids long with a predicted molecular mass of 27.6 and 27.7 kDa, respectively. VrCBF2 and VvCBF2 are 96% identical to each other and, respectively, 70 and 68% identical to VrCBF1. They are rich in serine (14.8/17%), alanine (9.6/10.3%), arginine (7.1/7.6%), proline (7.2/5.9%) and aspartic acid (6.4/6.3%), and have a pI of 9.89/9.71. Sequence type 3-specific primers amplified a 1.9 kb fragment on XbaI-digested and self-ligated genomic DNA. The subsequently amplified coding sequences of CBF3 from V. riparia and V. vinifera (Fig. 1) showed that VrCBF3 and VvCBF3 contained an open reading frame for proteins of 239 amino acids with a predicted molecular mass of 25.9 and 26 kDa, respectively. VrCBF3 and VvCBF3 are 96% identical to each other and 68/69 and 62/63% identical to VrCBF1 and VrCBF2. They are rich in serine (17.2/17.2%), alanine (10/10.5%), arginine (7.5/7.1%), leucine (6.7/6.7%) and glutamic acid (6.7/6.7%), and have a pI of 6.24/6.12. The grape CBF1, 2 and 3 proteins are very similar to A. thaliana CBF1 with an identity of 51, 46 and 42/44%, with as major differences an extra region of 20–29 amino acids within the C-terminal acidic domain and the presence

© 2006 The Authors Journal compilation © 2006 Blackwell Publishing Ltd, Plant, Cell and Environment, 29, 1410–1421

Stress-regulated grape CBFs 1415 of three serine repeats, one at the N-terminal region and the other two in the activation domain (Fig. 1). All Vitis CBF proteins contain an AP2 DNA-binding domain and characteristic CBF protein motifs (Jaglo et al. 2001; Owens et al. 2002; Al-daoud 2004) (Fig. 1). These include a putative NLS sequence, which is identical (HKRKAGRKKFRETRH) in Vitis CBF1 and 3, but has a slightly different sequence (HKRKTGR/KKKFRKTRH) in Vitis CBF2. However, the Vitis CBFs lack the proline (P) found at position one in the NLS of all other reported CBF proteins. A second sequence in Vitis CBF1 and 2 is identical to another consensus sequence in eudicot CBF proteins, DSAWRL, but slightly different in Vitis CBF3. In addition, an A motif with the Vitis CBF-specific sequence G/EDI/ VQV/FAAL/IxAA/TK/MAF and a P motif with the Vitis CBF-specific sequence MAEGLLLT/APP, but only part of the C-terminal LWSY domain, are present. Taken together, these results indicate that the isolated grape sequences are AtCBF orthologs, but have unique features.

The Vitis CBF gene family contains at least four members To determine how many similar CBF genes are present in grape, we hybridized the digested V. riparia and V. vinifera genomic DNA with three specific probes containing sequences corresponding to the region downstream of the AP2 domain of VrCBF1a, VrCBF2, and VrCBF3, and a non-specific probe containing the full-length VrCBF1a coding region. The Vitis CBF1-specific probe detected two bands in the EcoRI, HindIII and XbaI digest (Fig. 2b). The sizes of the fastest migrating HindIII and XbaI bands corresponded to the sizes expected for, respectively, CBF1a and CBF1b, based on the iPCR amplification of a 2.7 kb HindIII fragment and a 2.2 kb XbaI fragment with Vitis CBF1-specific primers. The Vitis CBF2-specific probe detected two bands each in EcoRI and HindIII, and one band in XbaI digests (Fig. 2c). One band was detected in all restriction digests with the Vitis CBF3-specific probe (Fig. 2d). The bands detected by the full-length VrCBF1a

(a) kb

(b)

(c)

probe included those detected by the Vitis CBF1- and Vitis CBF2-specific probes, however, not those detected by the Vitis CBF3-specific probe. This might be explained by two dissimilar regions in the sequence of Vitis CBF3 compared to the other two Vitis CBF genes (Fig. 1). The full-length VrCBF1a probe also detected an additional band in the EcoRI, HindIII and XbaI digests (indicated by arrowheads in Fig. 2a). The Southern blot analysis with V. vinifera genomic DNA showed the same pattern as V. riparia (Fig. 2a–d). These results indicate that there are at least four CBF genes in both V. riparia and V. vinifera: CBF1, CBF2, CBF3 and one unknown gene with high similarity to the CBF1 and CBF2 genes. Two different alleles appear to be present for CBF1, and possibly also for the other genes.

Vitis CBF1 localizes to the nucleus Transcription factors have to localize to the nucleus to activate genomic gene expression. The identified Vitis CBFs have a putative NLS sequence that is slightly different from that of all other reported CBF proteins, with a histidine instead of the bending proline at position 1 (Fig. 1). We therefore analysed, by agro-infiltration, the localization of a VrCBF1b–GFP fusion protein construct under the control of a CaMV35S promoter. The epidermal tissue of tobacco leaves was examined under the epifluorescence microscope and confocal microscope 2 d after agroinfiltration. The 35S::VrCBF1b–GFP construct induced fluorescence in nuclei only (Fig. 3a), whereas the 35S::GUS∆–GFP control showed GFP fluorescence throughout the whole cell (Fig. 3b). This result indicates that VrCBF1b can target to nuclei.

VrCBF1a, VrCBF1b and VvCBF1 transactivate expression of a CRT/DRE-containing gene If the grape CBF proteins function as transcription factors, then we would expect them to activate genes with a CRT/ DRE element in their promoter. We examined this for Vitis

(d)

Figure 2. Southern blot of Vitis riparia and Vitis vinifera DNA. The total genomic DNA (10 µg) was digested with restriction enzymes EcoRI (E), HindIII (H) and XbaI (X). The same membrane was hybridized consecutively with digoxygenin-labelled DNA probe corresponding to (a) the entire coding region of VrCBF1a (full-length probe), (b) the 3′-untranslated region of VrCBF1a (specific probe), (c) the 3′-untranslated region of VrCBF2 (specific probe) and (d) the 3′-untranslated region of VrCBF3 (specific probe). The arrow indicates the extra band detected by VrCBF1 fulllength probe, but not by the three specific probes. The results are representative of two independent experiments.

© 2006 The Authors Journal compilation © 2006 Blackwell Publishing Ltd, Plant, Cell and Environment, 29, 1410–1421

1416 H. Xiao et al.

(a)

(b)

(c)

(d)

Figure 3. Grape CBF protein localizes to the nucleus. Tobacco leaves were infiltrated with Agrobacterium containing plasmid with 35S::VrCBF1–mGFP5 or 35S::GUS–mGFP5 (control) and examined 2 d later. Confocal and corresponding differential interference contrast (DIC) images of tobacco epidermal cells showing nuclear localization of VrCBF1– green fluorescent protein (GFP) fusion protein (molecular mass ∼52 kDa) [ (a) and (c)] and cytoplasm localization of β-glucuronidase (GUS)–GFP fusion protein (molecular mass ∼53 kDa) [ (b) and (d)]. Nuclei are marked by arrows. The results are representative of two independent experiments.

CBF1 proteins by co-infiltration of tobacco leaves with two agrobacteria (Yang, Li & Qi 2000), one containing an effector plasmid with a CBF-encoding region driven by constitutive CaMV35S promoter and the other containing a reporter plasmid with a mGFP5–ER encoding region under the control of the cold-inducible RD29A promoter which includes four CRT/DRE cis-elements (YamaguchiShinozaki & Shinozaki 1993) (Fig. 4). Activation of expression from the RD29A promoter could therefore be detected as GFP fluorescence by epifluorescence microscopy and as GFP transcripts by RT–PCR. All cells in each field with on average 74 cells fluoresced with the 35S::mGFP5–ER positive control construct (7) (Fig. 4). Only one or two cells per field fluoresced with the RD29A::mGFP5–ER construct in the absence of CBF, suggesting that essentially no expression occurred [(5) and (6)] (Fig. 4). However, co-infiltration with the 35S::AtCBF control (1) (Fig. 4) resulted in an average of 20 fluorescent cells, and co-infiltration with the 35S::VrCBF1a (2) (Fig. 4), 35(S)::VrCBF1b (3) (Fig. 4) or 35S::VvCBF1 (4) (Fig. 4) gave on average 58, 52 or 18 cells with GFP fluorescence. The amount of GFP transcripts, as determined by RT–PCR, followed a similar pattern (Fig. 4b). This experiment was repeated two times with similar results. To confirm that the observed activation is caused by an interaction with CRT (as expected for a true CBF = CRTbinding factor), we also examined VrCBF1 transactivation of GFP or GUSPlus reporter behind four copies of CRT elements plus a minimal 35S promoter. The analysis of

RNA expression required the use of the intron-containing GUSPlus as reporter because this allowed the detection of spliced transcripts (i.e. plant-expressed RNA) and thereby avoid detection of some bacterial RNA that was expressed with the 35S minimal promoter constructs. GFP or GUSPlus expression analysis, as fluorescence or GUS activity and the presence of spliced GUS RNA transcripts, showed that VrCBF1 specifically interacts with CRT elements to cause an activation (Fig. 5). Taken together, these results indicate that Vitis CBF1 can activate the expression of a gene with CRT/DRE sequence as cis-element, presumably by binding to it. In addition, these results confirm that the Agrobacterium-mediated transient assay (Yang et al. 2000) is a simple and rapid method to analyse the in vivo interaction between transcription factors and cis-acting elements.

CBF1, CBF2 and CBF3 transcripts differentially accumulate in response to low temperature To investigate the cold induction of grape CBFs, grape plants were transferred to low temperature (4 °C) and leaves were collected after various periods of time. Semiquantitative RT–PCR with specific primers was used to analyse Vitis CBF expression. The identity of the amplified fragments was confirmed by sequencing representative

(a)

(b)

(c)

Figure 4. Transactivation of C-repeat (CRT)/dehydrationresponsive element (DRE)-containing promoter by VrCBF1, VvCBF1 and AtCBF1. (a) Combinations of reporter and effector plasmids used in the co-infiltration assays. The reporter GFP gene is driven by the RD29A promoter containing CRT/DRE elements, the effector plant CBF gene is driven by the cauliflower mosaic virus 35S (CaMV35S) promoter. (b) Semi-quantitative reverse transcription (RT)–PCR detection of GFP, CBF and ribulose 1·5bisphosphate carboxylase/oxygenase (Rubisco)-L (control) expression in tobacco leaves infiltrated with the reporter and effector plasmid combinations presented in (a). (c) Number of fluorescent cells observed by epifluorescent microscopy in epidermal tissue from tobacco leaves infiltrated with the reporter and effector plasmid combinations presented in (a). Data are mean and SD of 20 fields, each containing on average 74 cells, from at least two different infiltrated leaves. The results are representative of two independent experiments.

© 2006 The Authors Journal compilation © 2006 Blackwell Publishing Ltd, Plant, Cell and Environment, 29, 1410–1421

Stress-regulated grape CBFs 1417 (a)

(b)

(c) GUS activity (nm mg–1 min–1)

Reporter

GUSPlus AtCBF1 VrCBF1 MDH

Effector (1) Non–agroinfiltrated plants (2) min 35S:: GUSPlus (3) 4CRTmin 35S:: GUSPlus (4) 35S:: AtCBF1

+

min 35S:: GUSPlus

(5) 35S:: AtCBF1

+

4CRTmin 35S:: GUSPlus

(6) 35S:: VrCBF1

+

min 35S:: GUSPlus

(7) 35S:: VrCBF1

+

4CRTmin 35S: :GUSPlus

Figure 5. Transactivation of promoters with C-repeat (CRT) elements only. (a) Combinations of reporter and effector plasmids used in the co-infiltration assays. The GUSPlus reporter gene is driven by a promoter with or without four CRT elements in front of a minimal (−46) 35S promoter sequence, the effector plant CBF gene is driven by the constitutively expressing cauliflower mosaic virus 35S (CaMV35S) promoter. (b) Semi-quantitative reverse transcription (RT)–PCR detection of GUSPlus, CBF and MDH (control) expression in tobacco leaves infiltrated with the reporter and effector plasmid combinations presented in (a). (c) β-glucuronidase (GUS) activity (mean and SD) determined for tobacco leaves infiltrated with the reporter and activator combinations presented in (a). Similar expression results were obtained with green fluorescent protein (GFP) as reporter.

samples. Note that the Vitis CBF1-specific primers will detect transcripts from both CBF1a and CBF1b. Mature leaves of 2-month-old V. vinifera cv. Chardonnay plants under unstressed control conditions expressed VvCBF1, VvCBF2 and VvCBF3 (Fig. 6a, lane 1). The transcript levels were low for VvCBF2 and VvCBF3, but higher for VvCBF1, although to a varying extent in different experiments (compare Fig. 6a & d). Upon exposure to low temperature, the expression of VvCBF1 was detected for a short period, also if the initial pretreatment levels were very low, but decreased 1–2 h after the start of the treatment (Fig. 6a). Cold treatment enhanced the amount of VvCBF2 and VvCBF3 transcripts. The VvCBF2 transcript levels were highest during the 2–12 h period of treatment, then decreased but remained detectable to at least 7 d after transfer to 4 °C (Fig. 6a). The expression of VvCBF3 was barely detected during the first 24 h of exposure to 4 °C, whereas a significant increase was observed after 1–2 d, and this increase continued for at least a further 3 d. The expression profile of CBF1, CBF2 and CBF3 in response to cold in leaves from 2-month-old V. riparia Thunderbay leaves appeared to be very similar to that observed in V. vinifera cv. Chardonnay (Fig. 6a & b). Less-to-no transcripts were detected in leaves from plants that were not transferred to cold (Fig. 6c, control). In a similar experiment with mature leaves from 2-year-old V. vinifera cv. Chardonnay and V. riparia Thunderbay plants, only very low levels of CBF1 and CBF2 transcript were detected at any time point, although some expression of CBF3 was detected after 2 d of exposure to 4 °C (data not shown). A comparison of CBF expression between different tissues of the same 2year-old V. vinifera cv. Chardonnay plant showed that young leaves contained higher amounts of CBF1, CBF2 and CBF3 transcripts than mature leaves after cold induction (Fig. 6d). Low temperature enhanced all three transcripts in young leaves, apical tips, young buds and young

stems (Fig. 6e), except for CBF1 in leaves. Taken together, these results suggest that the younger the grape tissue, the higher the CBF1–3 expression level, and that CBF1, CBF2 and CBF3 expression is highest, respectively, immediately, after a few hours, or after a few days of starting a cold treatment.

Grape CBF transcripts accumulate in response to drought stress and ABA treatment Because young leaves expressed higher levels of CBF transcripts than mature leaves upon exposure to low temperature, we used detached young leaves from V. vinifera cv. Chardonnay plants for our next set of experiments. The leaves were placed on a filter paper with or without water to investigate the effect of drought stress, or on a filter paper with Tween solution without (control) or with 100 µM ABA to investigate the effect of exposure to ABA. Drying induced VvCBF1, VvCBF2 and VvCBF3 within 15 min to relatively high levels, but transcript levels declined later (Fig. 7a). VvCBF1 and VvCBF2, but not VvCBF3, were similarly induced by ABA (Fig. 7b). No transcripts were induced in water (not shown) and Tween controls (Fig. 7c).

DISCUSSION The results presented in this paper suggest the presence of at least five CBF-like genes in grape. Three of these genes were isolated from both V. riparia and V. vinifera and named, respectively, VrCBF1, VrCBF2, VrCBF3 and VvCBF1, VvCBF2, VvCBF3. We found two variants of VrCBF1 sequence, encoding proteins differing in only two amino acid residues. Because these two sequences were cloned in four independent experiments, we assume that they are not the result of PCR amplification or sequencing errors, but reflect the presence of two alleles in the grape

© 2006 The Authors Journal compilation © 2006 Blackwell Publishing Ltd, Plant, Cell and Environment, 29, 1410–1421

1418 H. Xiao et al. (a)

genome, although we cannot exclude the possibility that they represent two recently duplicated genes. Most likely, the two bands detected by the Vitis CBF1-specific probe on Southern blots correspond to the two sequences. The detection of two bands also by the Vitis CBF2-specific probe suggests that a similar situation might exist for this gene. A putative fourth gene was detected with the Vitis CBF1 fulllength probe, but not any of the Vitis CBF gene-specific probes. This fourth Vitis CBF-like gene still has to be isolated, but its detection under stringent hybridization conditions suggests that it is similar to Vitis CBF1 and CBF2. A fifth CBF gene in grape, with a shorter coding sequence, is represented by the 400 bp fragment that was amplified by the degenerate CBF-specific primers. The isolation and analysis of this gene and its expression are in progress and will be reported in a separate paper. The presence of a CBF gene family in plants appears common. Five to six CBF/ DREB1 genes have been reported for the eudicots A. thaliana and Brassica, at least 14 or more for rice and barley (Gilmour et al. 1998; Medina et al. 1999; Choi et al. 2002; Gao et al. 2002; Dubouzet et al. 2003; Xue 2003; Skinner et al. 2005). The fact that the CBF gene fragments were

(b)

(c)

(a)

(d)

(b)

(e)

(c)

Figure 6. Changes in Vitis CBF transcript levels in response to low temperature. Vitis vinifera cv. Chardonnay and Vitis riparia Thunderbay plants grown at 22 °C were subjected to low temperature (4 °C) treatment. Samples were taken at the indicated times, with zero time samples taken prior to treatment. The levels of CBF1, CBF2, CBF3 and MDH (control) transcripts were determined by semi-quantitative reverse transcription (RT)–PCR analysis. (a) Leaves from 2-month-old V. vinifera cv. Chardonnay. (b) Leaves from 2-month-old V. riparia Thunderbay. (c) Leaves from 2-month-old V. vinifera cv. Chardonnay and V. riparia Thunderbay kept at 22 °C (control treatment). (d) Apical tips, young leaves and mature leaves from 2-year-old V. vinifera cv. Chardonnay. (e) Apical tips, young leaves, young buds and young stems from 2-year-old V. vinifera cv. Chardonnay. The results are representative of at least two independent experiments.

Figure 7. Changes in Vitis CBF transcript levels in response to dehydration and abscisic acid (ABA). Detached young leaves of Vitis vinifera cv. Chardonnay plants were subjected to (a) drought, (b) ABA, (c) water plus Tween (water control) treatments, and the levels of CBF1, CBF2, CBF3 and MDH (control) transcripts were determined by semi-quantitative reverse transcription (RT)–PCR analysis. All zero time samples are of control plant tissues prior to treatments. The results are representative of two independent experiments.

© 2006 The Authors Journal compilation © 2006 Blackwell Publishing Ltd, Plant, Cell and Environment, 29, 1410–1421

Stress-regulated grape CBFs 1419 successfully isolated from grape by PCR with the degenerate primers CBFd1-H and CBFd4-C suggests that these two primers should be very useful in isolating additional CBF genes from other eudicots. To determine if the three Vitis CBF genes presented in this paper encode true CBFs, we analysed whether they are similar to eudicot CBFs with respect to: (1) CBF-specific amino acid domains; (2) induction of transcription from genes with CRT elements in their promoter; and (3) transcript accumulation in response to abiotic stress. The three Vitis CBF proteins, like all proteins belonging to the CBF gene family, have an AP2 domain and several other characteristic amino acid motifs (Jaglo et al. 2001; Al-daoud 2004). The sequence motifs are mostly as expected for eudicot CBF proteins (Al-daoud 2004), however, there are some notable differences. The first position of the putative nuclear localization sequence is a histidine, to give a stretch of four basic amino acids, instead of the proline present in all other eudicot CBFs reported so far. Our targeting and transactivation experiments suggest that the grape CBFs still function despite this change. Like the Prunus PaDREB1A and PaDREB1B, the Vitis CBFs lack a complete LWSY motif at the C-terminus. No function has yet been ascribed to this motif, so it is unknown what the consequence is of it missing. The Vitis CBFs do contain three serine repeats, one at the N-terminus, similar to several CBF proteins from Solanaceae (Jaglo et al. 2001), Rosaceae (Owens et al. 2002) and Poaceace (Choi et al. 2002; Dubouzet et al. 2003), and two in the activation domain, which seem to be unique (Fig. 1). These repeats might play a role in the activation of transcription by these proteins (Riechmann & Meyerowitz 1998). The grape CBF1 and CBF2, but not CBF3, proteins are basic. Basic CBF proteins are relatively unique for eudicot CBF proteins, only the Fragaria FaCBF1 and Prunus PaDREB1B were reported to be basic (see Al-daoud 2004). Finally, the three CBF proteins from both V. riparia and V. vinifera are larger than most of the other CBF proteins, 239–253 amino acid long, because of an extra region of 24–29 amino acids in their acidic domains. CBF proteins from A. thaliana bind to CRT/DRE elements to induce the expression of their target genes (JagloOttosen et al. 1998; Kasuga et al. 1999; Gilmour et al. 2000; Haake et al. 2002). We therefore investigated if a grape CBF can activate the expression of a gene with CRT elements in its promoter. Agro-infiltration experiments in tobacco leaves showed that Vitis CBF1 can activate reporter gene expression from promoters containing CRT elements as part of the RD29A promoter or in addition to a minimal 35S promoter (4CRT35S). This activation occurred at room temperature, showing that no coldinduced modification of the Vitis CBF is necessary. Recent preliminary freezing tests with transgenic A. thaliana lines show that over-expression of VrCBF1 results in increasing freezing tolerance compared to wild-type A. thaliana plants (data not shown). We conclude that the VrCBF1 and VvCBF1 reported here are most likely functional copies of the grape CBF gene family.

The three Vitis CBF genes were expected to respond to cold, like most CBF genes in other plants. Our results showed that CBF1, CBF2 and CBF3 transcripts were detectable in different tissues (apical tips, leaves, buds and stems) when exposed to low temperature, with higher levels in young tissues compared to mature tissues. It is possible that this reflects an adaptation to cold conditions in early spring when the buds burst and new leaves begin to develop. The three Vitis CBF transcripts accumulated in different time periods after exposure to low temperature: first CBF1 (minutes), then CBF2 (hours) and finally, after 2 d, CBF3 (days). Novillo et al. (2004) recently reported that the transcript levels of the A. thaliana CBF genes reflect the negative regulation of AtCBF1 and AtCBF3 expression by AtCBF2. Whether a similar situation exists for the grape CBF genes remains to be answered. Transcription of the three Vitis CBF genes was also induced by drought and ABA. Most CBF genes, including AtCBF1-3, are induced by cold but not drought (Gilmour et al. 1998; Shinwari et al. 1998; Medina et al. 1999), and only one gene, AtCBF4, is induced by drought but not cold (Haake et al. 2002). Conflicting reports have appeared regarding the responsiveness of AtCBF1-4 genes to ABA (Haake et al. 2002; Sakuma et al. 2002; Knight et al. 2004). Knight et al. (2004) argue that this is because the age and conditions of plants influence the response. ABA levels are thought to increase only transiently in response to low temperatures instead of cumulatively under drought conditions (Thomashow 1999). However, the transient relatively low levels of ABA might be sufficient to enhance the response to subsequent, stronger stresses. Indeed, Knight, Brandt & Knight (1998) showed in an earlier report that plant cells could have a memory of earlier stress encounters. The plants that gave an ABA response were smaller and therefore possibly more easily stressed and more sensitive to ABA (Knight et al. 2004). It is possible that a similar situation exists for the young grape tissues we used. This could explain the varying amounts of grape CBF1 transcripts we detected under control conditions. The recently reported response of at least one barley CBF gene to ABA confirms that not all CBF genes are ABA insensitive (Skinner et al. 2005). The results presented in this paper show that grape plants contain a CBF pathway, presumably to deal with abiotic stress. We did not present evidence to support the hypothesis that the CBF genes reported in this paper are responsible for the difference in freezing tolerance between V. riparia and V. vinifera. The Vitis CBF genes are induced in the same tissues, by the same cues and at the same time. However, an equal amount of transcripts does not necessarily mean equal stress tolerance. For example, CBF transcript levels were not correlated with a leaf orderdependent enhancement of freezing tolerance in coldacclimated A. thaliana (Takagi et al. 2003). It is also possible that the CBF expression of V. riparia and V. vinifera differs in a post-transcriptional step, for example, in their binding preference for CRT elements with different flanking sequences as shown for rice and barley CBF proteins

© 2006 The Authors Journal compilation © 2006 Blackwell Publishing Ltd, Plant, Cell and Environment, 29, 1410–1421

1420 H. Xiao et al. (Dubouzet et al. 2003; Xue 2003; Skinner et al. 2005). The observed higher induction by VrCBF1 compared to VvCBF1 in the transactivation assay (Fig. 4) might be caused by such a difference and deserves further study. Another possibility is that the situation in grape is similar to that shown for tomato (Zhang et al. 2004) in that the freezing-sensitive cultivar, V. vinifera, might have a smaller CBF regulon than the freezing-tolerant cultivar, V. riparia.

ACKNOWLEDGMENTS The research was supported by Chateau des Charmes Wineries, St. Catherine, Ontario, Canada and the Food Science Biotechnology Centre at the University of Guelph. We thank Sandra Stewart for technical assistance with preliminary experiments, Raymond Lee for critical reading of the manuscript and Judith Strommer for the supply of the cloning plasmid p35S–mGFP5ER.

REFERENCES Al-daoud F. (2004) Evolution of the CBF/DREB1 gene family into different types. MSc thesis, University of Guelph, Ontario, Canada. Choi D.-W., Rodriguez E.D. & Close T.J. (2002) Barley Cbf3 gene identification, expression pattern, and map location. Plant Physiology 129, 1781–1787. Dubouzet J.G., Sakuma Y., Ito Y., Kasuga M., Dubouzet E.G., Miura S., Seki M., Shinozaki K. & Yamaguchi-Shinozaki K. (2003) OsDREB genes in rice, Oryza sativa L., encode transcription activators that function in drought-, high salt-, and coldresponsive gene expression. Plant Journal 33, 751–763. Fowler S. & Thomashow M.F. (2002) Arabidopsis transcriptome profiling indicates that multiple regulatory pathways are activated during cold acclimation in addition to the CBF cold response pathway. Plant Cell 14, 1675–1690. Gao M.-J., Allard G., Byass L., Flanagan A.M. & Singh J. (2002) Regulation and characterization of four CBF transcription factors from Brassica napus. Plant Molecular Biology 49, 459–471. Gilmour S.J., Zarka D.G., Stockinger E.J., Salazar M.P., Houghton J.M. & Thomashow M.F. (1998) Low temperature regulation of the Arabidopsis CBF family of AP2 transcriptional activators as an early step in cold-induced COR gene expression. Plant Journal 16, 433–442. Gilmour S.J., Sebolt A.M., Salazar M.P., Everard J.D. & Thomashow M.F. (2000) Overexpression of the Arabidopsis CBF3 transcriptional activator mimics multiple biochemical changes associated with cold acclimation. Plant Physiology 124, 1854– 1865. Grebenok R.J., Peirson E., Lambert G.M., Gong F.-G., Afonso C.L., Halderman-Cahill R., Carrington J.C. & Galbraith D.W. (1997) Green-fluorescent protein fusions for efficient characterization of nuclear targeting. Plant Journal 11, 573–586. Haake V., Cook D., Riechmann J.L., Pineda O., Thomashow M.F. & Zhang J.Z. (2002) Transcription factor CBF4 is a regulator of drought adaptation in Arabidopsis. Plant Physiology 130, 639– 648. Haseloff J. & Siemering K.R. (1998) The uses of green fluorescent protein in plants. In Green Fluorescent Protein: Properties, Applications, and Protocols (eds M. Chalfie & S. Kain), pp. 191– 220. Wiley-Liss, New York, USA. Haseloff J., Siemering K.R., Prasher D.C. & Hodge S. (1997) Removal of a cryptic intron and subcellular localization of green

fluorescent protein are required to mark transgenic Arabidopsis plants brightly. Proceedings of the National Academy of Sciences of the USA 94, 2122–2127. Hofgen R. & Willmitzer L. (1988) Storage of competent cells for Agrobacterium transformation. Nucleic Acids Research 16, 9877. Jaglo K.R., Kleff S., Amundsen K.L., Zhang X., Haake V., Zhang J.Z., Deits T. & Thomashow M.F. (2001) Components of the Arabidopsis C-repeat/dehydration-responsive element binding factor cold-response pathway are conserved in Brassica napus and other plant species. Plant Physiology 127, 910–917. Jaglo-Ottosen K.R., Gilmour S.J., Zarka D.G., Schabenberger O. & Thomashow M.F. (1998) Arabidopsis CBF1 overexpression induces COR genes and enhances freezing tolerance. Science 280, 104–106. Jefferson R.A. (1987) Assaying chimeric genes in plants: the GUS gene fusion system. Plant Molecular Biology Reporter 5, 387– 405. Kasuga M., Liu Q., Miura S., Yamaguchi-Shinozaki K. & Shinozaki K. (1999) Improving plant drought, salt, and freezing tolerance by gene transfer of a single stress-inducible transcription factor. Nature Biotechnology 17, 287–291. Knight H., Brandt S. & Knight M.R. (1998) A history of stress alters drought calcium signalling pathways in Arabidopsis. Plant Journal 16, 681–687. Knight H., Zarka D.G., Okamoto H., Thomashow M.F. & Knight M.R. (2004) Abscisic acid induces CBF genes transcription and subsequent induction of cold-regulated genes via the CRT promoter element. Plant Physiology 135, 1710–1717. Lee R.W.H., Strommer J., Hodgins D., Shewen P.E., Niu Y. & Lo R.Y.C. (2001) Towards development of an edible vaccine against bovine pneumonic pasterellosis using transgenic white clover expressing a Mannheimia haemolytica A1 leukotoxin 50 fusion protein. Infection and Immunity 69, 5786–5793. Liu Q., Kasuga M., Sakuma Y., Abe H., Miura S., YamaguchiShinozaki K. & Shinozaki K. (1998) Two transcription factors, DREB1 and DREB2, with an EREBP/AP2 DNA binding domain separate two cellular signal transduction pathways in drought- and low-temperature-responsive gene expression, respectively, in Arabidopsis. Plant Cell 10, 1391–1406. Medina J., Bargues M., Terol J., Perez-Alonso M. & Salinas J. (1999) The Arabidopsis CBF gene family is composed of three genes encoding AP2 domain-containing proteins whose expression is regulated by low temperature but not by abscisic acid or dehydration. Plant Physiology 119, 463–469. Nassuth A., Pollari E., Helmeczy K., Stewart S. & Kofalvi S. (2000) Improved RNA extraction and one-tube RT–PCR assay for the simultaneous detection of control plant RNA plus several viruses in plant extracts. Journal of Virological Methods 90, 37–49. Novillo F., Alonso J.M., Ecker J.R. & Salinas J. (2004) CBF2/ DREB1C is a negative regulator of CBF1/DREB1B and CBF3/ DREB1A expression and plays a central role in stress tolerance in Arabidopsis. Proceedings of the National Academy of Sciences of the USA 101, 3985–3990. Owens C.L., Thomashow M.F., Hancock J.F. & Iezzoni A.F. (2002) CBF1 orthologs in sour cherry and strawberry and heterologous expression of CBF1 in strawberry. Journal of the American Society of Horticultural Sciences 127, 489–494. Riechmann J.L. & Meyerowitz E.M. (1998) The AP2/EREBP family of plant transcription factors. Journal of Biological Chemistry 379, 633–646. Sakuma Y., Liu Q., Dubouzet J.G., Abe H., Shinozaki K. & Yamaguchi-Shinozaki K. (2002) DNA-binding specificity of the ERF/AP2 domain of Arabidopsis DREBs, transcription factors involved in dehydration- and cold-inducible gene expression. Biochemical and Biophysical Research Communications 290, 998–1009.

© 2006 The Authors Journal compilation © 2006 Blackwell Publishing Ltd, Plant, Cell and Environment, 29, 1410–1421

Stress-regulated grape CBFs 1421 Sambrook J. & Russell D.N. (2001) Molecular Cloning: A Laboratory Manual, 3rd edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, USA. Seki M., Narusaka M., Ishida J., et al. (2002) Monitoring the expression profiles of 7000 Arabidopsis genes under drought, cold and high-salinity stresses using a full-length cDNA microarray. Plant Journal 31, 279–292. Shinwari Z.K., Nakashima K., Miura S., Kasuga M., Seki M., Yamaguchi-Shinozaki K. & Shinozaki K. (1998) An Arabidopsis gene family encoding DRE/CRT binding proteins involved in low-temperature–responsive gene expression. Biochemical and Biophysical Research Communications 250, 161–170. Skinner J.S., von Zitzewitz J., Szucs P., Marquez-Cedillo L., Filichkin T., Amundsen K., Stockinger E.J., Thomasow M.F., Chen T.N.N. & Hayes P.M. (2005) Structural, functional, and phylogenetic characterization of a large CBF gene family in barley. Plant Molecular Biology 59, 533–551. Stockinger E.J., Gilmour S.J. & Thomashow M.F. (1997) Arabidopsis thaliana CBF1 encodes an AP2 domain-containing transcriptional activator that binds to the C-repeat/DRE, a cis-acting DNA regulatory element that stimulates transcription in response to low temperature and water deficit. Proceedings of the National Academy of Sciences of the USA 94, 1035–1040. Takagi T., Nakamura M., Hayashi H., Inatsugi R., Yano R. & Nishida I. (2003) The leaf-order-dependent enhancement of freezing tolerance in cold-acclimated Arabidopsis rosettes is not

correlated with the transcript levels of the cold-inducible transcription factors of CBF/DREB1. Plant Cell Physiology 44, 922– 931. Thomashow M.F. (1999) Plant cold acclimation: freezing tolerance genes and regulatory mechanisms. Annual Review of Plant Physiology and Plant Molecular Biology 50, 571–599. Winkler A.J. (1970) General Viticulture. University of California, San Diego, CA, USA. Xue G.-P. (2003) The DNA-binding activity of an AP2 transcriptional activator HvCBF2 involved in regulation of lowtemperature responsive genes in barley is modulated by temperature. Plant Journal 33, 373–383. Yamaguchi-Shinozaki K. & Shinozaki K. (1993) Characterization of the expression of a desiccation-responsive rd29 gene of Arabidopsis thaliana and analysis of its promoter in transgenic plants. Molecular and General Genetics 263, 331–340. Yang Y., Li R. & Qi M. (2000) In vivo analysis of plant promoters and transcription factors by agroinfiltration of tobacco leaves. Plant Journal 22, 543–551. Zhang X., Fowler S.G., Cheng H., Lou Y., Rhee S.Y., Stockinger E.J. & Thomasow M.F. (2004) Freezing-sensitive tomato has a functional CBF cold-response pathway, but a CBF regulon that differs from that of freezing-tolerant Arabidopsis. Plant Journal 39, 905–919. Received 4 January 2006; accepted for publication 15 February 2006

© 2006 The Authors Journal compilation © 2006 Blackwell Publishing Ltd, Plant, Cell and Environment, 29, 1410–1421